Preprint
Article

This version is not peer-reviewed.

Novel Characterizations of the m-Generalized Group Inverse

Submitted:

07 October 2026

Posted:

09 October 2026

You are already at the latest version

Abstract
We study the m-generalized group inverse of Banach elements and establish several new properties. The main result is a novel characterization via the right hybrid (b,c)-inverse, revealing an intrinsic link between the two notions. Additive properties and image-based characterizations are also derived. These results extend the theory of the m-weak group inverse to a broader class of Banach elements and shed new light on infinite-dimensional operators.
Keywords: 
;  ;  ;  ;  

1. Introduction

Let A be a complex Banach algebra with a proper involution ∗. Here, the involution ∗ is proper if x * x = 0 ⟹ x = 0 for any x ∈ A . Let C n × n be the Banach algebra of all n × n complex matrices, with conjugate transpose ∗ as the involution. Then the involution ∗ is proper.
An element a ∈ A has group inverse provided that there exists x ∈ A such that
a x 2 = x , a x = x a , x a 2 = a .
Such x is unique if it exists, denoted by a # , and called the group inverse of a (see [1,15]).
In [26], Zou et al. introduced the weak group inverse inverse. An element a ∈ R has a weak group inverse if there exist x ∈ R and k ∈ N such that
a x 2 = x , ( a * a 2 x ) * = a * a 2 x , x a k + 1 = a k .
If such x exists, it is unique, and denote it by a Ⓦ (see [9,22,24,27]).
In [30], Zhou et al. introduced and studied the m-weak group inverse. An element a ∈ A has an m-weak group inverse if there exist x ∈ A and k ∈ N such that
a x 2 = x , ( a k ) * a m + 1 x = ( a k ) * a m , x a k + 1 = a k .
The preceding x is called the m-weak group inverse of a, and denoted by a Ⓦ m . The m-weak group inverse has been extensively studied from many different views, e.g., [5,10,12,13,18,19,20,21].
An element a ∈ A has g-Drazin inverse (i.e., generalized Drazin inverse) if there exists x ∈ A such that
a x 2 = x , a x = x a , a − a 2 x ∈ A q n i l .
Such x is unique, if it exists, and we denote it by a d . Here,
A q n i l = { x ∈ A | lim n → ∞ ‖ x n ‖ 1 n = 0 } .
Evidently, x ∈ A q n i l if and only if 1 + λ x ∈ A − 1 .
As a natural extension of the m-weak group inverse of complex matrices, with the aim of applying this notion to infinite-dimensional bounded linear operators on Banach spaces, we introduce and study a new generalized inverse (see [2]).
Definition 1.1.
An element a ∈ A has m-generalized group inverse if a ∈ A d and there exists x ∈ A such that
x = a x 2 , ( a d ) * a m + 1 x = ( a d ) * a m , lim n → ∞ | | a n − x a n + 1 | | 1 n = 0 .
The preceding x is called the m-generalized group inverse of a, and denoted by a ⓖ m .
This paper aims to investigate new properties of the m-generalized group inverse for elements in Banach *-algebras, thereby extending many of the properties of m-weak group inverse to broader classes, including infinite-dimensional bounded linear operators.
Let a , b , c ∈ A . Following Drazin [7], an element a ∈ R has right hybrid ( b , c ) -inverse if there exists an element x ∈ R such that
x a x = x , x R = b R and r ( x ) = r ( c ) .
If such x exists, it is unique and is denoted by a r h ( b , c ) . We refer the reader to [7,23,32] for more properties of the right hybrid ( b , c ) -inverse. In Section 2, we provide a new characterization of the m-generalized group inverse by exploring its relationship with the right hybrid ( b , c ) -inverse, establishing a vital connection between these generalized inverses.
In Section 3, We investigate the additive properties of the m-generalized group inverse. We establish necessary and sufficient conditions for the m-generalized group inverse of the sum a + b provided that 1 + a ⓖ m b is invertible.
Finally, in Section 4, we develop a novel characterization of the m-generalized group inverse by leveraging specific conditions involving its related images.
Throughout this paper, let A be a complex Banach *-algebra with a proper involution *. We denote by A − 1 , A d , A ⓖ , and A ⓖ m the sets of all invertible, generalized Drazin invertible, generalized group invertible, and m-generalized group invertible elements in A , respectively. Let N represent the set of all natural numbers. Let a ∈ A d . Let a π denote the spectral idempotent of a corresponding to { 0 } , i.e., a π = 1 − a a d . For an element x ∈ A , we use ℓ ( x ) and r ( x ) to denote the left and right annihilators of x, defined by ℓ ( x ) = { y ∈ A ∣ y x = 0 } and r ( x ) = { y ∈ A ∣ x y = 0 } , respectively. The image { x r r ∈ A } of x is given by i m ( x ) .

2. Characterizations via Right Hybrid ( b , c ) -Inverses

The aim of this section is to characterize the m-generalized group inverse by using the hybrid ( b , c ) -inverses. We begin with
Theorem 2.1.
Let a ∈ A and m ∈ N . Then the following are equivalent:
(1)
a ∈ A ⓖ m .
(2)
a ∈ A d and a ∈ A r h ( a d , ( a d ) * a m ) .
In this case, a ⓖ m = a r h ( a d , ( a d ) * a m ) .
Proof. 
( 1 ) ⇒ ( 2 ) In view of ???, a ∈ A d . By virtue of [2, Corollary 3.3], a ⓖ m = a ⓖ m a a ⓖ m and a ⓖ m = a d a a ⓖ m ∈ a d A . Moreover, we have a d = a ⓖ m a a d ∈ a ⓖ m A . This implies that a ⓖ m A = a d A . Let r ∈ r ( a ⓖ m ) . Then ( a d ) * a m r = ( a d ) * a m + 1 a ⓖ m r = 0 . Hence, r ( a ⓖ m ) ⊆ r ( a d ) * a . Let r ∈ r ( a d ) * a . Then ( a d ) * a r = 0 , and so ( a d ) * a d a 3 a ⓖ m r = ( a d ) * a 2 a ⓖ m r = ( a d ) * a r = 0 . Hence ( a d ) * a d a m + 2 a ⓖ m r = ( a d ) * a m + 1 a ⓖ m r = ( a d ) * a m r = 0 . This implies that a d a m + 2 a ⓖ m r = 0 . Thus a m + 1 a ⓖ m r = 0 . Thus a ⓖ m r = ( a ⓖ m ) 2 [ a 2 a ⓖ m r ] = 0 , and then r ( a d ) * a m ⊆ r ( a ⓖ m ) . Therefore r ( a ⓖ m ) = r ( a d ) * a m .
( 2 ) ⇒ ( 1 ) Set x = a r h ( a d , ( a d ) * a m ) . Then x a a d = a d , ( a d ) * a 2 x = ( a d ) * a , x ∈ a d A and r ( a d ) * a = r ( x ) . Then ( a d ) * a 2 x a d = ( a d ) * a a d . Since x ∈ a d A , we can find some z ∈ A such that a d z = x , and then ( a d ) * a d a 2 z a d = ( a d ) * a d a . Since the involution is proper, we deduce that a d a 2 z a d = a d a , and so a 2 x a d = a a d . Thus, x = a d z = ( a a d ) a d z = a 2 x ( a d ) 2 z . This implies that a a d x = x . Therefore a x 2 = a x a a d x = a a d x = x . Observe that
a n − x a n + 1 = a n − x a a d a n + 1 − x [ a n + 1 − a d a n + 2 ] = ( 1 − x a n ) ( a n − a d a n + 1 ) .
Accordingly,
lim n → ∞ | | a n − x a n + 1 | | 1 n = 0 .
Consequently, we have a ⓖ m = x , as asserted. □
Recall that an element a has the generalized group inverse if it has 1-generalized group inverse and we denote a ⓖ 1 by a ⓖ .
Corollary 2.2.
Let a ∈ A . Then the following are equivalent:
(1)
a ∈ A ⓖ .
(2)
a ∈ A d and a ∈ A r h ( a d , ( a d ) * a ) .
In this case, a ⓖ = a r h ( a d , ( a d ) * a ) .
Proof. 
This is obvious by choosing m = 1 . □
An element a ∈ A has generalized core-EP inverse if there exists x ∈ A such that
x = a x 2 , ( a x ) * = a x , lim n → ∞ | | a n − x a n + 1 | | 1 n = 0 .
Many properties of the generalized core-EP inverse can be found in [3,4].
Let a , b , c ∈ R . An element a has ( b , c ) -inverse provide that there exists x ∈ R such that
x a b = b , c a x = c and x ∈ b R x ⋂ x R c .
If such x exists, it is unique and denote it by a ( b , c ) (see [6,14,23]) for various properties of ( b , c ) -inverse.
Theorem 2.3.
Let a ∈ A ⓓ . Then
a ⓖ m = a ( a d , ( a ⓓ ) * a m ) .
Proof. 
Set x = a ⓖ m . In view of [2, Theorem 2.1], x = a m − 1 ( a m ) ⓖ . Then we have x = a m − 1 [ ( a m ) ⓓ ] 2 a m = a m − 1 [ a ⓓ ] 2 m a m = [ a ⓓ ] m + 1 a m . Then we check that
x = x ( a x ) = a a d a ⓖ a x ∈ a d A x , x = ( x a ) x = x m + 1 a m + 1 [ a ⓓ ] m + 1 a m = x m + 1 a a ⓓ a m = x m + 1 ( a a ⓓ ) * a m = x m + 1 ( a ⓓ a 2 a ⓓ ) * a m = x m + 1 ( a 2 a ⓓ ) * ( a ⓓ ) * a m ∈ x A ( a ⓓ ) * a m , x a a d = ( a ⓖ m a a d = a d , ( a ⓓ ) * a m a x = ( a ⓓ ) * a m + 1 [ a ⓓ ] m + 1 a m = ( a ⓓ ) * a a ⓓ a m = ( a ⓓ ) * ( a a ⓓ ) * a m = ( a ⓓ ) * a m .
Therefore a ⓖ m = a ( a d , ( a ⓓ ) * a m ) , as asserted. □
Corollary 2.4.
Let a ∈ A ⓓ . Then
a ⓖ = a ( a d , ( a ⓓ ) * a ) .
Proof. 
This is obvious by choosing m = 1 in Theorem 2.3. □
Now, we present the decompositions involving EP-like properties (see [6]).
Theorem 2.5.
Let a ∈ A . Then the following are equivalent:
(1)
a ∈ A ⓖ m .
(2)
There exist x , y ∈ A such that
a = x + y , x * a m − 1 y = y x = 0 , x ∈ A r h ( a d , ( a d ) a m ) # , y ∈ A q n i l .
Proof. 
( 1 ) ⇒ ( 2 ) Set z = a ⓖ m . Then we have
z = a z 2 , ( ( a m ) * a m + 1 z ) * = ( a m ) * a m + 1 z , lim n → ∞ | | a n − z a n + 1 | | 1 n = 0 .
In view of [2, Theorem 3.2], there exist x , y ∈ A such that
a = x + y , x * a m − 1 y = y x = 0 , x ∈ A # , y ∈ A q n i l .
Evidently, we have. x = a 2 z and y = a − a 2 z . Moreover, x has group inverse and z = x # = a ⓖ m . In view of Theorem 2.1, a ⓖ m = a r h ( a d , ( a d ) * a m ) . That is, we have
a ⓖ m = a ⓖ m a a ⓖ m , a ⓖ m A = a d A , r ( a ⓖ m ) = r ( a d ) * a m .
Therefore we verify that
a ⓖ m ( a 2 z ) a d = a ⓖ m a 2 a ⓖ m a d = a a ⓖ m a d = a d ; ( a d ) * a m ( a 2 z ) a ⓖ m = ( a d ) * a m a 2 ( a ⓖ m ) 2 = ( a d ) * a m + 1 a ⓖ m = ( a d ) * a m ; a ⓖ m = a a d a ⓖ m ∈ a d A .
If ( a d ) * a m r = 0 , then ( a d ) * a m + 1 a ⓖ m r = 0 . Hence ( a d ) * a d a m + 2 a ⓖ m r = 0 . Then a d a m + 2 a ⓖ m r = 0 . Thus a m + 1 a ⓖ m r = 0 . This implies that a ⓖ m r = 0 . This implies that r ( a d ) * a m ⊆ r a ⓖ m . Therefore x # = ( a 2 z ) r h ( a d , ( a d ) * a m ) = x r h ( a d , ( a d ) * a m ) . Accordingly, x ∈ A r h ( a d , ( a d ) * a m ) # , as required.
( 2 ) ⇒ ( 1 ) By hypothesis, there exist x , y ∈ A such that
a = x + y , x * a m − 1 y = y x = 0 , x ∈ A # , y ∈ A q n i l .
According to [2, Theorem 3.2], a ∈ A has m-generalized group inverse. □
We come now to characterize m-generalized group inverse by the polar-like property (see [4]).
Corollary 2.6.
Let a ∈ A . Then the following are equivalent:
(1)
a ∈ A ⓖ m .
(2)
a ∈ A d and there exists an idempotent p ∈ A such that
a + p ∈ A − 1 , 1 − p ∈ A r h ( a d , ( a d ) * a m ) , [ ( a m ) * a m p ] * = ( a m ) * a m p a n d p a = p a p ∈ A q n i l .
Proof. 
( 1 ) ⇒ ( 2 ) By virtue of Theorem 2.5, there exist z , y ∈ A such that
a = z + y , z * a m − 1 y = y z = 0 , z ∈ A r h ( a d , ( a d ) a m ) # , y ∈ A q n i l .
Let p = 1 − z z # . As in the proof of [2, Theorem 3.2], we verify that p = p 2 ∈ A and
a + p ∈ A − 1 , [ ( a m ) * a m p ] * = ( a m ) * a m p and p a = p a p ∈ A q n i l .
We will suffice to prove that 1 − p = z z # = a 2 ( a ⓖ m ) 2 = a a ⓖ m ∈ A r h ( a d , ( a d ) * a m ) . Clearly, we have
( a d ) * a m = ( a d ) * a m + 1 a ⓖ m = ( a d ) * a m + 1 a ⓖ m ( a d a 2 ) a ⓖ m ∈ ( a d ) * a m a a ⓖ m a d A .
If ( a d ) * a m a a ⓖ m a d r = 0 for some r ∈ A , then ( a d ) * a m a d r = 0 . Since the involution * is proper, we deduce that a d a m r = 0 , and so a d r = ( a d ) m ( a d a m ) r = 0 . This implies that r ∈ r ( a d ) . Hence, r ( a d ) * a m a a ⓖ m a d ⊆ r ( a d ) . By virtue of [7], a a ⓖ m ∈ A r h ( a d , ( a d ) * a m ) , as required.
( 2 ) ⇒ ( 1 ) This is proved by [2]. □

3. Additive Properties of m-Generalized Group Inverses

Let a ∈ A ⓖ m and b ∈ A . This section determines the necessary and sufficient conditions for the m-generalized group invertibility of a + b , provided that 1 + a ⓖ m b is invertible.
Lemma 3.1.
Let a ∈ A ⓖ m and b ∈ A . Then the following are equivalent:
(1)
( 1 − a a ⓖ m ) b = 0 .
(2)
a π b = 0 .
Proof. 
( 1 ) ⇒ ( 2 ) Since b = a a ⓖ m b , we deduce that a a d b = a a d a a ⓖ m ) b = a a ⓖ m ) b = b . Then a π b = 0 .
( 2 ) ⇒ ( 1 ) As b = a a d b , we have a a ⓖ m b = ( a a ⓖ m ) a a d b = a a d b = b . Therefore ( 1 − a a ⓖ m ) b = 0 , as asserted. □
Lemma 3.2.
Let a ∈ A ⓖ m and b ∈ A be such that 1 + a ⓖ m b ∈ A − 1 . Let f = a + b , f 0 = ( 1 + a ⓖ m b ) − 1 a ⓖ m and β = ( 1 + b a ⓖ m ) − 1 . Then the following are equivalent:
(1)
f f 0 2 = f 0 .
(2)
f f 0 = a a ⓖ m .
(3)
a π b a ⓖ m = 0 .
(4)
a π β a ⓖ m = 0 .
In this case, a π f d a ⓖ m = 0 .
Proof. 
Set α = ( 1 + a ⓖ m b ) − 1 . Obviously, we have α − 1 a ⓖ m = a ⓖ m β − 1 , we have f 0 = α a ⓖ m = a ⓖ m β .
( 1 ) ⇒ ( 4 ) Observe that f f 0 = ( a + b ) a ⓖ m β = a a ⓖ m β + b a ⓖ m β = a a ⓖ m β + 1 − β , we deduce that 1 − f f 0 = ( 1 − a a ⓖ m ) β . Since ( 1 − f f 0 ) f 0 = 0 , we deduce that ( 1 − a a ⓖ m ) β f 0 = 0 . Hence, ( 1 − a a ⓖ m ) β a ⓖ m β = 0 . As β ∈ A − 1 , we have ( 1 − a a ⓖ m ) β a ⓖ m = 0 , and then a π β a ⓖ m = 0 .
( 4 ) ⇒ ( 3 ) Since a π β a ⓖ m = 0 , we have
a π ( 1 + b a ⓖ m ) − 1 [ 1 + b a ⓖ m − b a ⓖ m ] a ⓖ m = 0 .
Hence,
a π a ⓖ m + a π [ − b a ⓖ m ] a ⓖ m = 0 .
This implies that a π b a ⓖ m = 0 , as required.
( 3 ) ⇒ ( 2 ) Since a π b a ⓖ m = 0 , we have a π ( 1 + b a ⓖ m ) = a π . This implies that a π β − 1 = a π , and then a π = a π β . That is, 1 − a a ⓖ m = ( 1 − a a ⓖ m ) β . Hence
f f 0 = ( a + b ) a ⓖ m β = a a ⓖ m β + b a ⓖ m β = a a ⓖ m β + ( 1 − β ) = 1 − 1 − a a ⓖ m ) β = 1 − ( 1 − a a ⓖ m ) = a a ⓖ m ,
as required.
( 2 ) ⇒ ( 1 ) As f f 0 = a a ⓖ m , we see that ( 1 − f f 0 ) f 0 = ( 1 − a a ⓖ m ) a ⓖ m β = 0 . This implies that f f 0 2 = f 0 , as asserted.
Since f f 0 2 = f 0 , we see that f d f f 0 = f 0 . Then we compute that
( 1 − a a ⓖ m ) f d a a ⓖ m = ( 1 − a a ⓖ m ) f d f f 0 = ( 1 − a a ⓖ m ) f 0 = ( 1 − a a ⓖ m ) ( f f 0 ) f 0 = ( 1 − a a ⓖ m ) ( a a ⓖ m ) f 0 = 0 .
□
Let f ∈ A and T ( f ) = { x ∈ A | a x 2 = x , lim n → ∞ | | a n − x a n + 1 | | 1 n = 0 } .
Theorem 3.3.
Let a ∈ A ⓖ m and b ∈ A be such that 1 + a ⓖ m b ∈ A − 1 . Let f = a + b and f 0 = ( 1 + a ⓖ m b ) − 1 a ⓖ m . Then the following are equivalent:
(1)
f ∈ A d and f 0 ∈ T ( f ) .
(2)
a π b a ⓖ m = 0 and a π f d ( 1 − a a ⓖ m ) = 0 .
(3)
a π b a ⓖ m = 0 and a π f d = 0 .
Proof. 
( 1 ) ⇒ ( 2 ) Since f 0 ∈ T ( f ) , we have f f 0 2 = f 0 . In view of Lemma 3.2, ( 1 − a a ⓖ m ) b a a ⓖ m = 0 . Then ( 1 − a a ⓖ m ) f a a ⓖ m = ( 1 − a a ⓖ m ) ( a + b ) a a ⓖ m = 0 . Since f a a ⓖ m = f 2 f 0 , we see that f a a ⓖ m ∈ A d . By virtue of [25, Lemma 2.2], ( 1 − a a ⓖ m ) f d a a ⓖ m = ( 1 − a a ⓖ m ) f a a ⓖ m d = 0 , as required.
( 2 ) ⇒ ( 3 ) In view of Lemma 3.1, we have ( 1 − a a ⓖ m ) f d a a ⓖ m . Since ( 1 − a a ⓖ m ) f d ( 1 − a a ⓖ m ) = 0 , we see that ( 1 − a a ⓖ m ) f d = 0 , as required.
( 3 ) ⇒ ( 1 ) Set α = ( 1 + a ⓖ m b ) − 1 . Obviously, f 0 f = α a ⓖ m ( a + b ) = α a ⓖ m a + α a ⓖ m b = α a ⓖ m a + α a ⓖ m a a ⓖ m b = α a ⓖ m a ( 1 + a ⓖ m b ) = α a ⓖ m a α − 1 . Then we observe that
( 1 − f 0 f ) f d = α ( 1 − a ⓖ m a ) α − 1 f d = α ( 1 − a ⓖ m a ) ( 1 + a ⓖ m b ) f d = α ( 1 − a ⓖ m a ) f d = α ( 1 − a ⓖ m a ) a a ⓖ m f d = 0 .
This implies that f d = f 0 f f d . We easily check that
f n − f 0 f n + 1 = f n − f d f n + 1 + f d f n + 1 − f 0 f n + 1 = f n − f d f n + 1 + f 0 f n + 2 f d − f 0 f n + 1 = f n − f d f n + 1 − f 0 f ( f n − f d f n + 1 ) = ( 1 − f 0 f ) ( f n − f d f n + 1 ) .
Hence,
| | f n − f 0 f n + 1 | | 1 n ≤ | | 1 − f 0 f | | 1 n | | f n − f d f n + 1 | | 1 n .
Thus we have
lim n → ∞ | | f n − f 0 f n + 1 | | 1 n = 0 .
In view of Lemma 3.2, f f 0 2 = f 0 . Therefore f 0 ∈ T ( f ) , as required. □
Lemma 3.4.
Let a ∈ A d . Then the following are equivalent:
(1)
a ∈ A ⓖ m .
(2)
( a d ) * a m ∈ ( a d ) * a d A .
(3)
There exists idempotent e ∈ A such that e A = a d A and e * a m ∈ e * e A .
Proof. 
( 1 ) ⇒ ( 2 ) This is proved by [2].
( 2 ) ⇒ ( 3 ) Choose e = a a d . Then e A = a d A . As ( a d ) * a m ∈ ( a d ) * a d A , we see that e * a m = ( a a d ) * a m ∈ a * ( a d ) * a d A = a * ( a d ) * a ( a d ) 2 A = ( a a d ) * ( a a d ) a d A ⊆ e * a m ∈ e * e A , as required.
( 3 ) ⇒ ( 1 ) Write a d = e y , e = a d z and e * a m = e * e t for some y , z , t ∈ A . Then ( a d ) * a m = y * ( e * a m ) = y * e * e t = ( e y ) * e t = ( a d ) * a d z t . Let x = z t . Then ( a d ) * a m = ( a d ) * a d x . Therefore a ∈ A ⓖ m by [2, Theorem 2.3]. □
We come now to the main result of this section.
Theorem 3.5.
Let a ∈ A ⓖ m and b ∈ A be such that 1 + a ⓖ m b ∈ A − 1 . Let f = a + b and f 0 = ( 1 + a ⓖ m b ) − 1 a ⓖ m . Then the following are equivalent:
(1)
f ∈ A ⓖ m and f ⓖ m = f 0 .
(2)
( a a ⓖ m ) * f m ( 1 − a a ⓖ m ) = 0 , a π b a ⓖ m = 0 and a π f d ( 1 − a a ⓖ m ) = 0 .
In this case, f f ⓖ m = a a ⓖ m .
Proof. 
( 1 ) ⇒ ( 2 ) Since f 0 = f ⓖ m ∈ T ( f ) , it follows by Theorem 3.3 that f f 0 = a a ⓖ m , a π b a ⓖ m = 0 and a π f d ( 1 − a a ⓖ m ) = 0 . Moreover, we derive that
( a a ⓖ m ) * f m ( 1 − a a ⓖ m ) = ( f f 0 ) * f m ( 1 − f f 0 ) = ( f d f f 0 ) * f m ( 1 − f f 0 ) = ( f f 0 ) * [ ( f d ) * f m − ( f d ) * f m + 1 f ⓖ m ] = 0 ,
as required.
( 2 ) ⇒ ( 1 ) By hypothesis, we have a π b a ⓖ m = 0 and a π f d ( 1 − a a ⓖ m ) = 0 . By virtue of Theorem 3.2, f 0 ∈ T ( f ) . In light of Lemma 3.2, f f 0 = a a ⓖ m . For any n ∈ N , we verify that ( 1 − f f 0 ) f d = ( 1 − f f 0 ) f n + 1 ( f d ) n + 2 = f ( f n − f 0 f n + 1 ) ( f d ) n + 2 = 0 , and then f f 0 f d = f d . Since ( a a ⓖ m ) * f m ( 1 − a a ⓖ m ) = 0 , we deduce that
( f d ) * f m ( 1 − f f 0 ) = ( f f 0 f d ) * f m ( 1 − f f 0 ) = ( a a ⓖ m f d ) * f m ( 1 − a a ⓖ m ) = ( f d ) * ( a a ⓖ m ) * f m ( 1 − a a ⓖ m ) = 0 .
Therefore ( f d ) * f m + 1 f 0 = ( f d ) * f m . Accordingly, f ⓖ m = f 0 , as asserted. □
Corollary 3.6.
Let a ∈ A ⓖ m and b ∈ A d be such that 1 + a ⓖ m b ∈ A − 1 . Let f = a + b and f 0 = ( 1 + a ⓖ m b ) − 1 a ⓖ m . If a b = b a and a * b = b a * , then the following are equivalent:
(1)
f ∈ A ⓖ and f ⓖ = f 0 .
(2)
( a a ⓖ ) * f m ( 1 − a a ⓖ ) = 0 and a π b d = 0 .
In this case, f ⓖ = f 0 = f d .
Proof. 
( 1 ) ⇒ ( 2 ) In view of Theorem 3.5, ( a a ⓖ m ) * f m ( 1 − a a ⓖ m ) = 0 and a π f d ( 1 − a a ⓖ m ) = 0 . By virtue of Lemma 3.1, ( 1 − a a ⓖ m ) f d ( 1 − a a ⓖ m ) = 0 . In light of ???, f d ( 1 − a a ⓖ m ) = ( 1 − a a ⓖ m ) f d . ( 1 − a a ⓖ m f d = 0 . By using Lemma 3.1 again, a π f d = 0 . According ???,
f d = ( a + b ) d = ( 1 + a d b ) d a d + b d ( 1 + a a π b d ) − 1 a π .
Thus
a π f d = a π [ ( 1 + a d b ) d a d + b d ( 1 + a a π b d ) − 1 a π ] = a π b d ( 1 + a a π b d ) − 1 = 0 .
This implies that a π b d = 0 , as required.
( 1 ) ⇒ ( 2 ) Since a π b d = 0 , a b = b a and a * b = b a * , we verify that a π b a ⓖ m = 0 and a π f d ( 1 − a a ⓖ m ) = 0 . Therefore we complete the proof by Theorem 3.5. □
We are ready to prove:
Theorem 3.7.
Let a ∈ A ⓖ m and b ∈ A . If f = a + b ∈ A ⓖ m , then the following are equivalent:
(1)
f f ⓖ m = a a ⓖ m .
(2)
a ⓖ m ( a + f ) f ⓖ m = a ⓖ m + f ⓖ m .
(3)
a ⓖ m ( f − a ) f ⓖ m = a ⓖ m − f ⓖ m .
(4)
1 + a ⓖ m b ∈ A − 1 , f ⓖ m = ( 1 + a ⓖ m b ) − 1 a ⓖ m .
Proof. 
( 1 ) ⇒ ( 2 ) Since a x = f y , we have x f y = x a x = x , and then
x a y = x a f y 2 = x a ( a x ) y = x a 2 x y = a x y = f y 2 = y .
Thus x ( a + f ) y = x a y + x f y = y + x , as desired.
( 2 ) ⇒ ( 1 ) Set x = a m ⓖ and y = f m ⓖ . Then x ( a + f ) y = x + y . Hence a x a x ( a + f ) y = a x a x + a x a y , and so a x ( a + f ) y = a x + a x a y . Thus a x f y = a x . On the other hand, we have x ( a + f ) y f 2 y = x f 2 y + y f 2 y . Hence x ( a + f ) f y = x f 2 y + f y , and so x a f y = f y . This implies that a x 2 a f y = a x f y , and then x a f y = a x f y . Accordingly, a x f y = f y . Therefore a x = a x f y = f y , as required.
( 1 ) ⇔ ( 3 ) This is proved by applying ( 1 ) ⇔ ( 2 ) to a and − f .
( 4 ) ⇒ ( 1 ) This is proved in Theorem 3.5.
( 3 ) ⇒ ( 4 ) Obviously, we have 1 + a ⓖ m b = 1 − a ⓖ m a + a ⓖ m f . For any n ∈ N , we see that
| | a a ⓖ m − a ⓖ m a 2 a ⓖ m | | 1 n = | | a n ( a ⓖ m ) n − a ⓖ m a n + 1 ( a ⓖ m ) n | | 1 n ≤ | | a n − a ⓖ m a n + 1 | | 1 n | | a ⓖ m | | .
Therefore
lim n → ∞ | | a a ⓖ m − a ⓖ m a 2 a ⓖ m | | 1 n = 0 .
Hence a a ⓖ m = a ⓖ m a 2 a ⓖ m . As f f ⓖ m = a a ⓖ m , we see that
( 1 − a ⓖ m a ) f ⓖ m = ( 1 − a ⓖ m a ) ( f f ⓖ m ) f ⓖ m = ( 1 − a ⓖ m a ) ( a a ⓖ m ) f ⓖ m = [ a a ⓖ m − a ⓖ m a 2 a ⓖ m ] f ⓖ m = 0
and a ⓖ m f f ⓖ m b = a ⓖ m b . Then we verify that
[ 1 + a ⓖ m b ] [ 1 − f ⓖ m b ] = [ 1 − a ⓖ m a + a ⓖ m f ] [ 1 − f ⓖ m b ] = 1 − a ⓖ m a + a ⓖ m f − a ⓖ m b = 1 .
Moreover, we see that
( 1 − f ⓖ m f ) a ⓖ m b = ( 1 − f ⓖ m f ) ( a a ⓖ m ) a ⓖ m b = ( 1 − f ⓖ m f ) ( f f ⓖ m ) a ⓖ m b = 0
and f ⓖ m a a ⓖ m b = f ⓖ m b . Then we check that
[ 1 − f ⓖ m b ] [ 1 + a ⓖ m b ] = [ 1 − f ⓖ m f + f ⓖ m a ] [ 1 + a ⓖ m b ] = 1 − f ⓖ m f + f ⓖ m a + f ⓖ m a a ⓖ m b = 1 − f ⓖ m f + f ⓖ m a + f ⓖ m b = 1 .
Hence, 1 + a ⓖ m b ∈ A − 1 . Clearly, we have
( 1 + a ⓖ m b ) f ⓖ m = [ 1 + a ⓖ m ( f − a ) ] f ⓖ m = f ⓖ m + a ⓖ m ( f − a ) ] f ⓖ m = f ⓖ m + ( a ⓖ m − f ⓖ m ) = a ⓖ m .
Therefore f ⓖ m = ( 1 + a ⓖ m b ) − 1 a ⓖ m , as asserted. □
Corollary 3.8.
Let a ∈ A ⓖ m and b ∈ A . If f = a + b ∈ A ⓖ , then the following are equivalent:
(1)
f f ⓖ = a a ⓖ .
(2)
a ⓖ ( a + f ) f ⓖ = a ⓖ + f ⓖ .
(3)
a ⓖ ( f − a ) f ⓖ = a ⓖ − f ⓖ .
(4)
1 + a ⓖ b ∈ A − 1 , f ⓖ = ( 1 + a ⓖ b ) − 1 a ⓖ .
Proof. 
This is obvious by choosing m = 1 in Theorem 3.7. □
Example 3.9.
Let H = ℓ 2 ( N ) be the Hilbert space of square-summable complex sequences, with the standard orthonormal basis { e n } n = 1 ∞ where e n = ( 0 , … , 0 , 1 , 0 , … ) .
We define T : H → H as a diagonal operator acting on the first two coordinates:
T ( x 1 , x 2 , x 3 , … ) = ( x 1 , 2 x 2 , 0 , 0 , … ) .
The group inverse of T, denoted T # , is given by:
T # ( y 1 , y 2 , y 3 , … ) = y 1 , 1 2 y 2 , 0 , 0 , … .
Since T is self-adjoint, T * ( y ) = ( y 1 , 2 y 2 , 0 , … ) .
We define S : H → H as a rank-two operator that mixes coordinates:
S ( x 1 , x 2 , x 3 , x 4 , … ) = x 1 + x 3 , 2 x 2 + 2 x 4 , 0 , 0 , … .
The group inverse S # is given by:
S # ( x 1 , x 2 , x 3 , x 4 , … ) = 0 , 0 , x 1 , 1 2 x 2 , 0 , … .
Consequently, we have S S # = T T # = P span { e 1 , e 2 } .
We define V : H → H as a weighted shift operator acting on the tail of the sequence:
V ( x 1 , x 2 , x 3 , x 4 , … ) = 0 , 0 , 1 2 x 3 , 1 3 x 4 , … .
The spectral radius r ( V ) = 0 as ∥ V n ∥ 1 / n → 0 .
The range of S and T is contained in span { e 1 , e 2 } . Since V e 1 = 0 and V e 2 = 0 , we have V ( S x ) = 0 and V ( T x ) = 0 .
The range of V is contained in span { e 3 , e 4 , … } . Since S * and T * map vectors into span { e 1 , e 2 } , and V x is orthogonal to this subspace, we have S * ( V x ) = 0 and T * ( V x ) = 0 . We easily check that
  • T and S have group inverses.
  • V is quasinilpotent.
  • T T # = S S # = P span { e 1 , e 2 } .
  • V S = V T = 0 and S * V = T * V = 0 .
Let α = T + V and γ = S + V . Since V is quasinilpotent and commutes appropriately with the projections (i.e., T T # V = V T T # = 0 ), α and γ have generalized group inverses. Specifically, the group inverse of α is α ⓖ = T # . We easily check that α α ⓖ = ( T + V ) T # = T T # + V T # = T T # . Likewise, we have γ γ ⓖ = S S # . Hence α α ⓖ = γ γ ⓖ .
Set β = γ − α . Then:
β ( x 1 , x 2 , x 3 , … ) = ( S + V ) − ( T + V ) = S − T = S ( x 1 , x 2 , x 3 , … ) − T ( x 1 , x 2 , x 3 , … ) = ( x 1 + x 3 , 2 x 2 + 2 x 4 , 0 , … ) − ( x 1 , 2 x 2 , 0 , … ) = ( x 3 , 2 x 4 , 0 , 0 , … ) .
Then M : = I + α ⓖ β = I + T # β . For any vector x = ( x 1 , x 2 , x 3 , x 4 , … ) :
M ( x ) = ( x 1 , x 2 , x 3 , x 4 , … ) + ( x 3 , x 4 , 0 , … ) = ( x 1 + x 3 , x 2 + x 4 , x 3 , x 4 , … ) .
The matrix representation of M with respect to the basis { e n } is:
M = I + T # β = 1 0 1 0 0 ⋯ 0 1 0 1 0 ⋯ 0 0 1 0 0 ⋯ 0 0 0 1 0 ⋯ ⋮ ⋮ ⋮ ⋮ ⋱ ⋮
This is an upper-triangular matrix with all diagonal entries equal to 1. Therefore, the operator M is invertible. Evidently, we have ( I + α ⓖ β ) − 1 = I − T # β . Explicitly, the action is given by:
( I + α ⓖ β ) − 1 ( x 1 , x 2 , x 3 , x 4 , … ) = ( x 1 − x 3 , x 2 − x 4 , x 3 , x 4 , … ) .
Moreover, we directly verify that γ ⓖ = ( I + α ⓖ β ) − 1 α ⓖ , as required.

4. Equivalent Characterizations Involving Images

In this section, we present distinct characterizations based on specific conditions involving their related images. We now derive
Theorem 4.1.
Let a ∈ A and m ∈ N . Then the following are equivalent:
(1)
a ∈ A ⓖ m .
(2)
A = i m ( a d ) + r ( a d ) * a m .
(3)
A = i m ( a d ) ⨁ r ( a d ) * a m .
(4)
i m ( a * a m ) = ( a * a m ) i m ( a d ) + ( a * a m ) r ( a d ) * a m .
(5)
i m ( a * a m ) = ( a * a m ) i m ( a d ) ⨁ ( a * a m ) r ( a d ) * a m .
Proof. 
( 1 ) ⇒ ( 2 ) Clearly, a a ⓖ m = a d ( a 2 a ⓖ m ) and ( a d ) * a m ( 1 − a a ⓖ m ) = ( a d ) * a m − ( a d ) * a m + 1 a ⓖ m = 0 . Then 1 = a a ⓖ m + ( 1 − a a ⓖ m ) ∈ i m ( a d ) + r ( a d ) * a m , as desired.
( 2 ) ⇒ ( 3 ) If r ∈ i m ( a d ) ⋂ r ( a d ) * a m , then we may write r = a d s for some s ∈ A . Hence ( a d ) a m a d s = 0 . As the involution is proper, we have a d a m s = 0 . This implies that r = a d s = ( a d ) m ( a d a m s ) = 0 . Thus, i m ( a d ) ⋂ r ( a d ) * a m = 0 , as desired.
( 3 ) ⇒ ( 5 ) ⇒ ( 4 ) These are obvious.
( 4 ) ⇒ ( 1 ) Write a * a m = a * a m x + a * a m y , where x ∈ a d A and y ∈ ∈ r ( a d ) * a m . Then ( a d ) * a m = ( a d ) * a m x + ( a d ) * a m y = ( a d ) * a m x . WE infer that ( a d ) * a m ∈ ( a d ) * a d A . Therefore we complete the proof by Lemma 3.4. □
Corollary 4.2.
Let a ∈ A . Then the following are equivalent:
(1)
a ∈ A ⓖ .
(2)
A = i m ( a d ) ⨁ r ( a d ) * a .
(3)
A = i m ( a d ) + r ( a d ) * a .
(4)
i m ( a * a ) = ( a * a ) i m ( a d ) ⨁ ( a * a ) r ( a d ) * a .
(5)
i m ( a * a ) = ( a * a ) i m ( a d ) + ( a * a ) r ( a d ) * a .
Proof. 
This is obvious by choosing m = 1 in Theorem 4.1. □
We are ready to prove:
Theorem 4.3.
Let a ∈ A d and m ∈ N . Then the following are equivalent:
(1)
a ∈ A ⓖ m .
(2)
There exists x ∈ A such that x a x = x , i m ( x ) = i m ( a d ) , i m ( a m ) * a d ⊆ i m ( x * ) .
(3)
There exists x ∈ A such that x a x = x , i m ( x ) = i m ( a d ) , ℓ ( x * ) ⊆ ℓ ( a m ) * a d .
(4)
There exists x ∈ A such that x a x = x , ℓ ( x ) ⊆ ℓ ( a d ) , ℓ ( x * ) ⊆ ℓ ( a m ) * a d .
In this case, a ⓖ m = x .
Proof. 
( 1 ) ⇒ ( 2 ) Set x = a m ⓖ . As in the proof of Theorem 2.5, we have x a x = x , x A = a d A . Since ( a d ) * a m + 1 x = ( a d ) * a m , we see that ( a m ) * a d = x * ( a m + 1 ) * a d ∈ x * A , as required.
( 2 ) ⇒ ( 3 ) ⇒ ( 4 ) These are trivial.
( 4 ) ⇒ ( 1 ) Since x a x = x , we have x a − 1 ∈ ℓ ( x ) ⊆ ℓ ( a d ) . Hence, x a a d = a d . Then for any n ∈ N , we have
a n − x a n + 1 = a n − x a ( a n − a d a n + 1 ) + ( x a a d ) a n + 1 = ( 1 − x a ) ( a n − a d a n + 1 ) .
This implies that
lim n → ∞ | | a n − x a n + 1 | | 1 n = 0 .
As x a a d = a d , we get a x a a d = a a d , and then a x a d = a d . This implies that a x − 1 ∈ ℓ ( a d ) ℓ ( x ) . Hence, ( a x − 1 ) x = 0 , and so a x 2 = a x . Clearly, x * a * x * = x * , and so x * a * − 1 ∈ ℓ ( x * ) ⊆ ℓ ( a m ) * a d . Thus ( x * a * − 1 ) ( a m ) * a d = 0 . Therefore ( a d ) * a m + 1 x = ( a d ) * a m , as required. □
Corollary 4.4.
Let a ∈ A d and m ∈ N . Then the following are equivalent:
(1)
a ∈ A ⓖ m .
(2)
There exists an idempotent e ∈ A such that i m ( e ) = i m ( a d ) and i m ( e * ) ⊆ i m ( a m ) * a d .
(3)
There exists an idempotent e ∈ A such that ℓ ( e ) = ℓ ( a d ) and r ( e ) ⊆ r ( a d ) * a m .
In this case, a ⓖ m = x .
Proof. 
( 1 ) ⇒ ( 2 ) Choose e = a a m ⓖ . Then e = a d a 2 a m ⓖ , a d = e a d . Then i m ( e ) = i m ( a d ) . Since ( a d ) * a m = ( a d ) * a m + 1 a m ⓖ = ( a d ) * a m e , we deduce that A ( a d ) * a m ⊆ A e , as required.
( 2 ) ⇒ ( 3 ) This is obvious.
( 3 ) ⇒ ( 1 ) Choose x = a d e . Then 1 − a a d ∈ ℓ ( a d ) ⊆ ℓ ( e ) . Hence, e = a a d e . If r ∈ ℓ ( x ) , then r a d e = 0 , then r a d ∈ ℓ ( e ) ⊆ ℓ ( a d ) . Hence r ( a d ) 2 = 0 , and so r ( a d ) = 0 . Thus ℓ ( x ) ⊆ ℓ ( a d ) .
As r ( e ) ⊆ r ( a d ) * a m , we see that ( a d ) * a m ( 1 − e ) = 0 . This implies that ( a d ) * a m = ( a d ) * a m e = ( a d ) * a m + 1 x . Accordingly, x * ( a m + 1 ) * a d = ( a m ) * a d . We infer that ℓ ( x * ) ⊆ ℓ ( a m ) * a d .
Moreover, we verify that
x a x = ( a d e ) a ( a d e ) = a d e ( a a d e ) = a d e 2 = a d e = x .
Therefore a ⓖ m = x by Theorem 4.3. □
Corollary 4.5.
Let a ∈ A d . Then the following are equivalent:
(1)
a ∈ A ⓖ .
(2)
There exists x ∈ A such that x a x = x , i m ( x ) = i m ( a d ) , i m ( a * a d ) ⊆ i m ( x * ) .
(3)
There exists x ∈ A such that x a x = x , i m ( x ) = i m ( a d ) , ℓ ( x * ) ⊆ ℓ a * a d .
(4)
There exists x ∈ A such that x a x = x , ℓ ( x ) ⊆ ℓ ( a d ) , ℓ ( x * ) ⊆ ℓ a * a d .
(5)
There exists an idempotent e ∈ A such that i m ( e ) = i m ( a d ) and i m ( e * ) ⊆ i m ( a * a d ) .
(6)
There exists an idempotent e ∈ A such that ℓ ( e ) = ℓ ( a d ) and r ( e ) ⊆ r ( a d ) * a .
In this case, a ⓖ m = x .
Proof. 
This is clear by choosing m = 1 in Theorem 4.3 and Corollary 4.4. □
Let a ∈ A and E ( a ) = { a x ∈ A | a x 2 = x , lim n → ∞ | | a n − x a n + 1 | | 1 n = 0 } .
Theorem 4.6.
Let a ∈ A d and m ∈ N . Then the following are equivalent:
(1)
a ∈ A ⓖ m .
(2)
There exists e ∈ E ( a ) such that ( a m ) * a m e * = ( a m ) * a m e .
(3)
There exists e ∈ E ( a ) such that ( a d ) * a m e = ( a d ) * a m .
(4)
There exists an idempotent e ∈ i m ( a d ) such that ( a d ) * a m e = ( a d ) * a m .
Proof. 
( 1 ) ⇒ ( 2 ) Set e = a a ⓖ m . Then e ∈ E ( a ) and ( a m ) * a m e * = ( a m ) * a m e by [2].
( 2 ) ⇒ ( 1 ) Write e = a x for some x ∈ T ( a ) . Then a x 2 = x and lim n → ∞ | | a n − x a n + 1 | | 1 n = 0 } . Clearly, ( a m ) * a m e = ( a m ) * a m + 1 x , and then ( a m ) * a m + 1 x * = ( a m ) * a m + 1 x . Therefore a ∈ A ⓖ m by [2].
( 1 ) ⇒ ( 3 ) Choose e = a a ⓖ m . Then e ∈ E ( a ) and ( a d ) * a m e = ( a d ) * a m , as required.
( 3 ) ⇒ ( 1 ) This is obvious.
( 1 ) ⇒ ( 4 ) Set e = a a ⓖ m . Then e 2 = e ∈ a a d A . By the argument above, we have ( a d ) * a m e = ( a d ) * a m .
( 4 ) ⇒ ( 1 ) Write e = a a d x for some x ∈ A . Then ( a d ) * a m = ( a d ) * a m ( a a d x ) = ( a d ) * a d ( a m + 1 x ) ∈ ( a d ) * a d A . Therefore we complete the proof by Lemma 3.4. □
As an immediate consequence, we derive
Corollary 4.7.
Let a ∈ A d . Then the following are equivalent:
(1)
a ∈ A ⓖ .
(2)
There exists e ∈ E ( a ) such that a * a e * = a * a e .
(3)
There exists e ∈ E ( a ) such that ( a d ) * a e = ( a d ) * a .
(4)
There exists e ∈ i m ( a d ) such that ( a d ) * a e = ( a d ) * a .

Disclosure Statements

No potential conflict of interest was reported by the authors.

References

  1. Boasso, E.; Kantún-Montiel, G. The (b,c)-inverse in rings and in the Banach context. Mediterr. J. Math. 2017, 14. [Google Scholar] [CrossRef]
  2. Chen, H. On m-generalized group inverse in Banach *-algebras. Mediterr. J. Math. 2025, 22. [Google Scholar] [CrossRef]
  3. Chen, H.; Sheibani, M. Generalized weighted core inverse in Banach *-algebras. Filomat 2024, 38, 3691–3706. [Google Scholar] [CrossRef]
  4. Chen, H.; Sheibani, M. Properties of generalized weighted core inverses in Banach *-algebras. J. Algebra Appl. 2025, 2550358. [Google Scholar] [CrossRef]
  5. Chen, Y.; Yuan, L.; Zhang, X. New characterizations and representations of the weak group inverse. Filomat 2024, 38, 6703–6713. [Google Scholar] [CrossRef]
  6. Drazin, M.P. A class of outer generalized inverses. Linear Algebra Appl. 2012, 436, 1909–1923. [Google Scholar] [CrossRef]
  7. Drazin, M.P. Hybrid (b,c)-inverses and five finiteness properties in rings, semigroups and categories. Comm. Algebra 2021, 49, 2265–2277. [Google Scholar] [CrossRef]
  8. Drazin, M.P. EP-like properties of (b,c)-inverses. Linear Algebra Appl. 2023, 669, 136–149. [Google Scholar] [CrossRef]
  9. Ferreyra, D.E.; Orquera, V.; Thome, N. A weak group inverse for rectangular matrices. Rev. R. Acad. Cienc. Exactas Fis.Nat. Ser. A Mat. 2019, 113, 3727–3740. [Google Scholar] [CrossRef]
  10. Gao, J.; Zu, K.; Wang, Q. A. M-Weak Group Inverse Rectangular Matrices. 2312.10704v1 [math.R.A] 17 Dec 2023.
  11. Gao, Y.; Chen, J. Pseudo core inverses in rings with involution. Comm. Algebra 2018, 46, 38–50. [Google Scholar] [CrossRef]
  12. Jiang, W.; Zuo, K. Further characterizations of the m-weak group inverse of a complex matrix. AIMS Math. 2020, 7, 17369–17392. [Google Scholar]
  13. Li, W.; Chen, J.; Zhou, Y. Characterizations and representations of weak core inverses and m-weak group inverses. Turk. J. Math. 2023, 47, 1453–1468. [Google Scholar] [CrossRef]
  14. Ke, Y.; Gao, Y.; Chen, J. Representations of the (b,c)-inverses in rings with involution. Filomat 2017, 31, 2867–2875. [Google Scholar] [CrossRef]
  15. Mihajlovic, N. Group inverse and core inverse in Banach and C*-algebras. Comm. Algebra 2020, 48, 1803–1818. [Google Scholar] [CrossRef]
  16. Mosić, D. Core-EP inverses in Banach algebras. Linear Multilinear Algebra 2021, 69, 2976–2989. [Google Scholar] [CrossRef]
  17. Mosić, D.; Staninirovic, P.S. Representations for the weak group inverse. Appl. Math. Comput. Article ID 125957, 19 p. 2021, 397. [Google Scholar] [CrossRef]
  18. Mosić, D.; Zhang, D. New representations and properties of the m-weak group inverse. Result. Math. 2023, 78. [Google Scholar] [CrossRef]
  19. Mosić, D.; Staninirovic, P.S.; Kazakovtsev, L.A. Application of m-weak group inverse in solving optimization problems. Rev. R. Acad. Cienc. Exactas Fis. Nat. Ser. A Mat. 2024, 118(No. 1, Paper No. 13), 21 p. [Google Scholar] [CrossRef]
  20. Mosić, D.; Stanimirović, P.S.; Kazakovtsev, L.A. Minimization problem solvable by weighted m-weak group inverse. J. Appl. Math. Comput. 2024, 70, 6259–6281. [Google Scholar] [CrossRef]
  21. Mosić, D.; Zhang, D. New representations and properties of the m-weak group inverse. Result. Math. 2023, 78. [Google Scholar] [CrossRef]
  22. Wang, H.; Chen, J. Weak group inverse. Open Math. 2018, 16, 1218–1232. [Google Scholar] [CrossRef]
  23. Wang, L. Characterizations of the (b. C)-Inverse A Ring. 1507.01446v1 [math.RA] 4 Jun 2015. [CrossRef]
  24. Yan, H.; Wang, H.; Zuo, K.; Chen, Y. Further characterizations of the weak group inverse of matrices and the weak group matrix. AIMS Math. 2021, 6, 9322–9341. [Google Scholar] [CrossRef]
  25. Zhang, D.; Mosić, D. Explicit formulae for the generalized Drazin inverse of block matrices over a Banach algebra. Filomat 2018, 32, 5907–5917. [Google Scholar] [CrossRef]
  26. Zhou, M.; Chen, J.; Zhou, Y. Weak group inverses in proper *-rings. J. Algebra Appl. Article ID 2050238, 14 p. 2020, 19. [Google Scholar] [CrossRef]
  27. Zhou, M.; Chen, J.; Zhou, Y.; Thome, N. Weak group inverses and partial isometries in proper *-rings. Linear Multilinear Algebra 2021, 70, 1–16. [Google Scholar] [CrossRef]
  28. Zhou, Y. Generalized Inverses of Sums and Products of Elements. Ph.D. Thesis, Southeast University, Nanjing, 2024. [Google Scholar]
  29. Zhou, Y.; Chen, J.; Thome, N. Additive properties and absorption laws for generalized inverses. Linear Multilinear Algebra 2026, 74, 213–227. [Google Scholar] [CrossRef]
  30. Zhou, Y.; Chen, J.; Zhou, M. m-weak group inverses in a ring with involution. Rev. R. Acad. Cienc. Exactas Fís. Nat. Ser. A Mat. 2021, 115 Paper(No. 2), 12 p. [Google Scholar]
  31. Zhu, H.; Patricio, P. Characterizations for pseudo core inverses in a ring with involution. Linear Multilinear Algebra 2019, 67, 1109–1120. [Google Scholar] [CrossRef]
  32. Zhu, H.; Wu, L.; Peng, F.; et al. Characterizations and representations of left and right hybrid (b,c)-inverses in rings. Rev. Real. Acad. Cienc. Exactas Fis. Nat. Ser. A-Mat. 2021, 22. [Google Scholar] [CrossRef]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
Copyright: This open access article is published under a Creative Commons CC BY 4.0 license, which permit the free download, distribution, and reuse, provided that the author and preprint are cited in any reuse.